merging_datasets

Overview

merging background and presence data into one giant dataset, and timesliced subsets

A note to anyone who might happen to stumble across this… I am a beginner in R and have had no exposure to similar languages. I don’t know what I’m doing. The code herein is unlikely to be elegant and there are probably more efficient ways of running the code.

Built with ‘r getRversion()’.

Package dependencies

You can load them using the following code which uses a function called ipak. Note this function checks to see if the packages are installed first.

packages <- c("plyr") 
source("./src/ipak.R")
ipak(packages)

plyr 
TRUE 

read in the presence and background

presencemerging <- read.csv("../output/bio/presence_points_without_envdata_relooped_glbathy_nodup.csv", header = TRUE)
backgroundmerging <- read.csv("../output/bio/background_complete_obs_cels_globot_cellid_xyzt.csv", header = TRUE)

nrow(test200301a)

[1] 5228

now check

colnames(presencemerging)

 [1] "cell_id"                "year"                   "month"                  "depthlayerno"           "id"                     "decimalLatitude"        "decimalLongitude"      
 [8] "datecollected"          "institutioncode"        "individualcount"        "depth"                  "resname"                "originalscientificname" "collectioncode"        
[15] "day"                    "occurrence"             "nafo_zone"              "gear"                   "longitude_meters"       "latitude_meters"        "amo_sample"            
[22] "amo_prev"               "amo_winter"             "depth_layer"            "bottom_depth"           "total_cell_obs"         "yymm_cell_obs"          "chl_surface"           
[29] "chl_depth"              "mlp_surface"            "o2_surface"             "o2_depth"               "salinity_surface"       "salinity_depth"         "ssh_surface"           
[36] "temp_surface"           "temp_depth"             "nao_sample"             "nao_prev"               "nao_winter"             "XXtotal_cell_obs_xyzt"  "temp_celsius_depth"    
[43] "temp_celsius_surface"   "longitude_meters.1"     "latitude_meters.1"      "bottom_depth_glorys"    "longitude_meters.2"     "latitude_meters.2"      "cell_id_3d"            
[50] "cell_id_xyzt"           "total_cell_obs_xyzt"   

colnames(backgroundmerging)

 [1] "cell_id"                "year"                   "month"                  "depthlayerno"           "X"                      "longitude_meters"       "latitude_meters"       
 [8] "decimalLongitude"       "decimalLatitude"        "nafo_zone"              "id"                     "originalscientificname" "amo_sample"             "amo_prev"              
[15] "amo_winter"             "nao_sample"             "nao_prev"               "nao_winter"             "depth_layer"            "bottom_depth"           "total_cell_obs_xy"     
[22] "total_cell_obs_xyt"     "total_cell_obs_xyzt"    "temp_depth"             "temp_surface"           "salinity_depth"         "salinity_surface"       "chl_depth"             
[29] "chl_surface"            "o2_depth"               "o2_surface"             "mlp_surface"            "ssh_surface"            "longitude_meters.1"     "latitude_meters.1"     
[36] "optional"               "occurrence"             "temp_celsius_depth"     "temp_celsius_surface"   "bottom_depth_glorys"    "longitude_meters.2"     "latitude_meters.2"     
[43] "optional.1"             "cell_id_3d"             "cell_id_xyzt"          

ok so the two dataframed need a spot of cleaning

presencemerging <- subset(presencemerging, select = -c(longitude_meters.1, latitude_meters.1, longitude_meters.2, latitude_meters.2, XXtotal_cell_obs_xyzt))

Error in eval(substitute(select), nl, parent.frame()) : 
  object 'longitude_meters.1' not found

for consistency, rename a couple of presencemerging columns

names(presencemerging)[names(presencemerging)=="total_cell_obs"] <- "total_cell_obs_xy"
names(presencemerging)[names(presencemerging)=="yymm_cell_obs"] <- "total_cell_obs_xyt"

now merge the two datasets

presab <- rbind.fill(presencemerging, backgroundmerging)
write.csv(presab, "../output/bio/presab.csv", row.names = FALSE)
head(presab)

lovely jubley

weighted 10k for each month

For each month, the 10,000 background points will be garnered from background points created for each month-year and weighted to the observations. For example, for January 22% of the unique observations by cell occurred in 2003, 49% in 2004, 7% in 2005, and 21% in 2006. Thus, 22% of the background points for the January model will come from 2003, 49% from 2004, 7% in 2005, and 21% in 2006. See ‘Percentage of occurrences by unique cell (xyzt)’ in Appendix Observations per Cell in results document + excel + data_exploration.

Some other analysis has indicated that background points have been gathered from 1BCDEF (ouside Canada EEZ and inside Greenland EEZ). Just do a table here

nrow(nafo1)

[1] 139235

and the spread of these funny nafo points in year month…

nafo1_yymm <- with(nafo1, table(year, month))
write.csv(nafo1_yymm, file = "../output/bio/background_nafo1x_mth_yr.csv")
nafo1_yymm

      month
year     1   2   3   4   5   6   7   8   9  10  11  12
  1998   0   0   0 932 848 897 936   0 914 952 896 868
  1999   0   0 878 909 910 900 885   0 895 950 889 911
  2000   0   0 968 925 910 908 907   0 958 922 903 916
  2001   0 902 941 963 935 933 965   0 906 871 923 914
  2002   0   0 905 929 918 905 922   0 941 903 906 872
  2003   0   0 864 873 895 910 935   0 913 878 926 907
  2004 902   0   0 889 934 869 927 924 968 932 904 897
  2005 962 865 900 962 877 917 890 901 907 862 872 933
  2006 895 866 926 958 939 898 917 932 914 892 931 881
  2007   0   0 943 902 895 886 947 908 885 920 857 928
  2008   0   0 914 894 931 921 944 913 906 919 934 869
  2009   0   0 867 921 875 835 928 899 974 933 923 912
  2010   0   0 906 869 901 902 875 937 897 851 931 867
  2011   0   0   0 956 924 875 924 946 940 888 904 897
  2012   0   0   0   0   0 921 929 918 890   0   0   0
  2013   0   0   0   0   0   0 949 892 924 876 935   0
  2014   0   0   0   0   0   0 865 897 953 941   0   0
  2015   0   0   0   0   0   0   0 876 921   0 890   0

ok so nothing too crazy… lets remove these nafo 1x points from the dataset

nafounique <- unique(backno1$nafo_zone)
nafounique

 [1] 3Ps          0A           0B           2H           2J           3O           3L           2G           4Vn          3K           4S           3N           4Vs         
[14] 4W           3M           4R           5Y           HudsonStrait 4T           4X           5Ze          3Pn         
Levels: 0A 0B 1B 1C 1D 1E 1F 2G 2H 2J 3K 3L 3M 3N 3O 3Pn 3Ps 4R 4S 4T 4Vn 4Vs 4W 4X 5Y 5Ze HudsonStrait

ok now randomly select background points from each timeslice as per this table (see results doc appendix observations per cell + excel spreadsheet)

    1   2   3   4   5   6   7   8   9   10  11  12

1998 0 0 0 846 790 835 518 0 358 1315 1123 107 1999 0 0 1433 603 595 968 648 0 620 169 835 551 2000 0 0 1277 603 537 746 492 0 652 590 1025 725 2001 0 5000 748 990 751 637 389 0 390 405 495 1430 2002 0 0 996 1062 761 676 440 0 811 624 420 966 2003 2222 0 717 1004 888 809 492 0 318 590 311 1014 2004 4938 0 0 732 790 797 674 1004 652 472 604 1101 2005 700 4000 872 316 644 676 1010 1270 556 776 858 416 2006 2140 1000 1558 301 49 790 674 1410 588 1096 616 783 2007 0 0 623 746 508 593 1813 1404 636 624 553 821 2008 0 0 934 516 741 797 777 1418 564 455 748 589 2009 0 0 312 946 1405 351 207 582 580 556 915 454 2010 0 0 530 560 663 720 648 596 747 1096 875 560 2011 0 0 0 775 878 605 285 758 533 1062 593 483 2012 0 0 0 0 0 0 208 793 604 0 0 0 2013 0 0 0 0 0 0 699 660 509 119 12 0 2014 0 0 0 0 0 0 26 77 818 51 0 0 2015 0 0 0 0 0 0 0 28 64 0 17 0 Total 10000 10000 10000 10000 10000 10000 10000 10000 10000 10000 10000 10000

Ok so for Jan: 2003: 2222 XXXXX THIS IS MISSING!!!! 2004: 4938 2005: 700 2006: 2140

nrow(jan2004)

[1] 4938

for feb 2001: 5000 2005: 4000 2006: 1000

feb2001 <- subset(backno1, year == "2001" & month == "2")
feb2001 <- sample_n(feb2001, 5000) 
feb2005 <- subset(backno1, year == "2005" & month == "2")
feb2005 <- sample_n(feb2005, 4000) 
feb2006 <- subset(backno1, year == "2006" & month == "2")
feb2006 <- sample_n(feb2006, 1000) 
feb <- rbind(feb2001, feb2005, feb2006)

for march 1999 1433 2000 1277 2001 748 2002 996 2003 717 2005 872 2006 1558 2007 623 2008 934 2009 312 2010 530

mar1999 <- subset(backno1, year == "1999" & month == "3")
mar1999 <- sample_n(mar1999, 1433) 
mar2000 <- subset(backno1, year == "2000" & month == "3")
mar2000 <- sample_n(mar2000, 1277) 
mar2001 <- subset(backno1, year == "2001" & month == "3")
mar2001 <- sample_n(mar2001, 748)
mar2002 <- subset(backno1, year == "2002" & month == "3")
mar2002 <- sample_n(mar2002, 996)
mar2003 <- subset(backno1, year == "2003" & month == "3")
mar2003 <- sample_n(mar2003, 717)
mar2005 <- subset(backno1, year == "2005" & month == "3")
mar2005 <- sample_n(mar2005, 872)
mar2006 <- subset(backno1, year == "2006" & month == "3")
mar2006 <- sample_n(mar2006, 1558)
mar2007 <- subset(backno1, year == "2007" & month == "3")
mar2007 <- sample_n(mar2007, 623)
mar2008 <- subset(backno1, year == "2008" & month == "3")
mar2008 <- sample_n(mar2008, 934)
mar2009 <- subset(backno1, year == "2009" & month == "3")
mar2009 <- sample_n(mar2009, 312)
mar2010 <- subset(backno1, year == "2010" & month == "3")
mar2010 <- sample_n(mar2010, 530)
mar <- rbind(mar1999, mar2000, mar2001, mar2002, mar2003, mar2005, mar2006, mar2007, mar2008, mar2009, mar2010)

For april 1998 846 1999 603 2000 603 2001 990 2002 1062 2003 1004 2004 732 2005 316 2006 301 2007 746 2008 516 2009 946 2010 560 2011 775

apr1998 <- subset(backno1, year == "1998" & month == "4")
apr1998 <- sample_n(apr1998, 846) 
apr1999 <- subset(backno1, year == "1999" & month == "4")
apr1999 <- sample_n(apr1999, 603) 
apr2000 <- subset(backno1, year == "2000" & month == "4")
apr2000 <- sample_n(apr2000, 603) 
apr2001 <- subset(backno1, year == "2001" & month == "4")
apr2001 <- sample_n(apr2001, 990)
apr2002 <- subset(backno1, year == "2002" & month == "4")
apr2002 <- sample_n(apr2002, 1062)
apr2003 <- subset(backno1, year == "2003" & month == "4")
apr2003 <- sample_n(apr2003, 1004)
apr2004 <- subset(backno1, year == "2004" & month == "4")
apr2004 <- sample_n(apr2004, 732)
apr2005 <- subset(backno1, year == "2005" & month == "4")
apr2005 <- sample_n(apr2005, 316)
apr2006 <- subset(backno1, year == "2006" & month == "4")
apr2006 <- sample_n(apr2006, 301)
apr2007 <- subset(backno1, year == "2007" & month == "4")
apr2007 <- sample_n(apr2007, 746)
apr2008 <- subset(backno1, year == "2008" & month == "4")
apr2008 <- sample_n(apr2008, 516)
apr2009 <- subset(backno1, year == "2009" & month == "4")
apr2009 <- sample_n(apr2009, 946)
apr2010 <- subset(backno1, year == "2010" & month == "4")
apr2010 <- sample_n(apr2010, 560)
apr2011 <- subset(backno1, year == "2011" & month == "4")
apr2011 <- sample_n(apr2011, 775)
apr <- rbind(apr1998, apr1999, apr2000, apr2001, apr2002, apr2003, apr2004, apr2005, apr2006, apr2007, apr2008, apr2009, apr2010, apr2011)

For May 1998 790 1999 595 2000 537 2001 751 2002 761 2003 888 2004 790 2005 644 2006 49 2007 508 2008 741 2009 1405 2010 663 2011 878

may1998 <- subset(backno1, year == "1998" & month == "5")
may1998 <- sample_n(may1998, 790) 
may1999 <- subset(backno1, year == "1999" & month == "5")
may1999 <- sample_n(may1999, 595) 
may2000 <- subset(backno1, year == "2000" & month == "5")
may2000 <- sample_n(may2000, 537) 
may2001 <- subset(backno1, year == "2001" & month == "5")
may2001 <- sample_n(may2001, 751)
may2002 <- subset(backno1, year == "2002" & month == "5")
may2002 <- sample_n(may2002, 761)
may2003 <- subset(backno1, year == "2003" & month == "5")
may2003 <- sample_n(may2003, 888)
may2004 <- subset(backno1, year == "2004" & month == "5")
may2004 <- sample_n(may2004, 790)
may2005 <- subset(backno1, year == "2005" & month == "5")
may2005 <- sample_n(may2005, 644)
may2006 <- subset(backno1, year == "2006" & month == "5")
may2006 <- sample_n(may2006, 49)
may2007 <- subset(backno1, year == "2007" & month == "5")
may2007 <- sample_n(may2007, 508)
may2008 <- subset(backno1, year == "2008" & month == "5")
may2008 <- sample_n(may2008, 741)
may2009 <- subset(backno1, year == "2009" & month == "5")
may2009 <- sample_n(may2009, 1405)
may2010 <- subset(backno1, year == "2010" & month == "5")
may2010 <- sample_n(may2010, 663)
may2011 <- subset(backno1, year == "2011" & month == "5")
may2011 <- sample_n(may2011, 878)
may <- rbind(may1998, may1999, may2000, may2001, may2002, may2003, may2004, may2005, may2006, may2007, may2008, may2009, may2010, may2011)

for June

1998 835 1999 968 2000 746 2001 637 2002 676 2003 809 2004 797 2005 676 2006 790 2007 593 2008 797 2009 351 2010 720 2011 605

jun1998 <- subset(backno1, year == "1998" & month == "6")
jun1998 <- sample_n(jun1998, 835) 
jun1999 <- subset(backno1, year == "1999" & month == "6")
jun1999 <- sample_n(jun1999, 968) 
jun2000 <- subset(backno1, year == "2000" & month == "6")
jun2000 <- sample_n(jun2000, 746) 
jun2001 <- subset(backno1, year == "2001" & month == "6")
jun2001 <- sample_n(jun2001, 637)
jun2002 <- subset(backno1, year == "2002" & month == "6")
jun2002 <- sample_n(jun2002, 676)
jun2003 <- subset(backno1, year == "2003" & month == "6")
jun2003 <- sample_n(jun2003, 809)
jun2004 <- subset(backno1, year == "2004" & month == "6")
jun2004 <- sample_n(jun2004, 797)
jun2005 <- subset(backno1, year == "2005" & month == "6")
jun2005 <- sample_n(jun2005, 676)
jun2006 <- subset(backno1, year == "2006" & month == "6")
jun2006 <- sample_n(jun2006, 790)
jun2007 <- subset(backno1, year == "2007" & month == "6")
jun2007 <- sample_n(jun2007, 593)
jun2008 <- subset(backno1, year == "2008" & month == "6")
jun2008 <- sample_n(jun2008, 797)
jun2009 <- subset(backno1, year == "2009" & month == "6")
jun2009 <- sample_n(jun2009, 351)
jun2010 <- subset(backno1, year == "2010" & month == "6")
jun2010 <- sample_n(jun2010, 720)
jun2011 <- subset(backno1, year == "2011" & month == "6")
jun2011 <- sample_n(jun2011, 605)
jun <- rbind(jun1998, jun1999, jun2000, jun2001, jun2002, jun2003, jun2004, jun2005, jun2006, jun2007, jun2008, jun2009, jun2010, jun2011)

July

1998 518 1999 648 2000 492 2001 389 2002 440 2003 492 2004 674 2005 1010 2006 674 2007 1813 2008 777 2009 207 2010 648 2011 285 2012 208 2013 699 2014 26

jul1998 <- subset(backno1, year == "1998" & month == "7")
jul1998 <- sample_n(jul1998, 518) 
jul1999 <- subset(backno1, year == "1999" & month == "7")
jul1999 <- sample_n(jul1999, 648) 
jul2000 <- subset(backno1, year == "2000" & month == "7")
jul2000 <- sample_n(jul2000, 492) 
jul2001 <- subset(backno1, year == "2001" & month == "7")
jul2001 <- sample_n(jul2001, 389)
jul2002 <- subset(backno1, year == "2002" & month == "7")
jul2002 <- sample_n(jul2002, 440)
jul2003 <- subset(backno1, year == "2003" & month == "7")
jul2003 <- sample_n(jul2003, 492)
jul2004 <- subset(backno1, year == "2004" & month == "7")
jul2004 <- sample_n(jul2004, 674)
jul2005 <- subset(backno1, year == "2005" & month == "7")
jul2005 <- sample_n(jul2005, 1010)
jul2006 <- subset(backno1, year == "2006" & month == "7")
jul2006 <- sample_n(jul2006, 674)
jul2007 <- subset(backno1, year == "2007" & month == "7")
jul2007 <- sample_n(jul2007, 1813)
jul2008 <- subset(backno1, year == "2008" & month == "7")
jul2008 <- sample_n(jul2008, 777)
jul2009 <- subset(backno1, year == "2009" & month == "7")
jul2009 <- sample_n(jul2009, 207)
jul2010 <- subset(backno1, year == "2010" & month == "7")
jul2010 <- sample_n(jul2010, 648)
jul2011 <- subset(backno1, year == "2011" & month == "7")
jul2011 <- sample_n(jul2011, 285)
jul2012 <- subset(backno1, year == "2012" & month == "7")
jul2012 <- sample_n(jul2012, 208)
jul2013 <- subset(backno1, year == "2013" & month == "7")
jul2013 <- sample_n(jul2013, 699)
jul2014 <- subset(backno1, year == "2014" & month == "7")
jul2014 <- sample_n(jul2014, 26)
jul <- rbind(jul1998, jul1999, jul2000, jul2001, jul2002, jul2003, jul2004, jul2005, jul2006, jul2007, jul2008, jul2009, jul2010, jul2011, jul2012, jul2013, jul2014)

For Aug 2004 1004 2005 1270 2006 1410 2007 1404 2008 1418 2009 582 2010 596 2011 758 2012 793 2013 660 2014 77 2015 28

aug2004 <- subset(backno1, year == "2004" & month == "8")
aug2004 <- sample_n(aug2004, 1004)
aug2005 <- subset(backno1, year == "2005" & month == "8")
aug2005 <- sample_n(aug2005, 1270)
aug2006 <- subset(backno1, year == "2006" & month == "8")
aug2006 <- sample_n(aug2006, 1410)
aug2007 <- subset(backno1, year == "2007" & month == "8")
aug2007 <- sample_n(aug2007, 1404)
aug2008 <- subset(backno1, year == "2008" & month == "8")
aug2008 <- sample_n(aug2008, 1418)
aug2009 <- subset(backno1, year == "2009" & month == "8")
aug2009 <- sample_n(aug2009, 582)
aug2010 <- subset(backno1, year == "2010" & month == "8")
aug2010 <- sample_n(aug2010, 596)
aug2011 <- subset(backno1, year == "2011" & month == "8")
aug2011 <- sample_n(aug2011, 758)
aug2012 <- subset(backno1, year == "2012" & month == "8")
aug2012 <- sample_n(aug2012, 793)
aug2013 <- subset(backno1, year == "2013" & month == "8")
aug2013 <- sample_n(aug2013, 660)
aug2014 <- subset(backno1, year == "2014" & month == "8")
aug2014 <- sample_n(aug2014, 77)
aug2015 <- subset(backno1, year == "2015" & month == "8")
aug2015 <- sample_n(aug2015, 28)
aug <- rbind(aug2004, aug2005, aug2006, aug2007, aug2008, aug2009, aug2010, aug2011, aug2012, aug2013, aug2014, aug2015)

For sep 1998 358 1999 620 2000 652 2001 390 2002 811 2003 318 2004 652 2005 556 2006 588 2007 636 2008 564 2009 580 2010 747 2011 533 2012 604 2013 509 2014 818 2015 64

sep1998 <- subset(backno1, year == "1998" & month == "9")
sep1998 <- sample_n(sep1998, 358) 
sep1999 <- subset(backno1, year == "1999" & month == "9")
sep1999 <- sample_n(sep1999, 620) 
sep2000 <- subset(backno1, year == "2000" & month == "9")
sep2000 <- sample_n(sep2000, 652) 
sep2001 <- subset(backno1, year == "2001" & month == "9")
sep2001 <- sample_n(sep2001, 390)
sep2002 <- subset(backno1, year == "2002" & month == "9")
sep2002 <- sample_n(sep2002, 811)
sep2003 <- subset(backno1, year == "2003" & month == "9")
sep2003 <- sample_n(sep2003, 318)
sep2004 <- subset(backno1, year == "2004" & month == "9")
sep2004 <- sample_n(sep2004, 652)
sep2005 <- subset(backno1, year == "2005" & month == "9")
sep2005 <- sample_n(sep2005, 556)
sep2006 <- subset(backno1, year == "2006" & month == "9")
sep2006 <- sample_n(sep2006, 588)
sep2007 <- subset(backno1, year == "2007" & month == "9")
sep2007 <- sample_n(sep2007, 636)
sep2008 <- subset(backno1, year == "2008" & month == "9")
sep2008 <- sample_n(sep2008, 564)
sep2009 <- subset(backno1, year == "2009" & month == "9")
sep2009 <- sample_n(sep2009, 580)
sep2010 <- subset(backno1, year == "2010" & month == "9")
sep2010 <- sample_n(sep2010, 747)
sep2011 <- subset(backno1, year == "2011" & month == "9")
sep2011 <- sample_n(sep2011, 533)
sep2012 <- subset(backno1, year == "2012" & month == "9")
sep2012 <- sample_n(sep2012, 604)
sep2013 <- subset(backno1, year == "2013" & month == "9")
sep2013 <- sample_n(sep2013, 509)
sep2014 <- subset(backno1, year == "2014" & month == "9")
sep2014 <- sample_n(sep2014, 818)
sep2015 <- subset(backno1, year == "2015" & month == "9")
sep2015 <- sample_n(sep2015, 64)
sep <- rbind(sep1998, sep1999, sep2000, sep2001, sep2002, sep2003, sep2004, sep2005, sep2006, sep2007, sep2008, sep2009, sep2010, sep2011, sep2012, sep2013, sep2014, sep2015)

For oct 1998 1315 1999 169 2000 590 2001 405 2002 624 2003 590 2004 472 2005 776 2006 1096 2007 624 2008 455 2009 556 2010 1096 2011 1062 2013 119 2014 51

oct1998 <- subset(backno1, year == "1998" & month == "10")
oct1998 <- sample_n(oct1998, 1315) 
oct1999 <- subset(backno1, year == "1999" & month == "10")
oct1999 <- sample_n(oct1999, 169) 
oct2000 <- subset(backno1, year == "2000" & month == "10")
oct2000 <- sample_n(oct2000, 590) 
oct2001 <- subset(backno1, year == "2001" & month == "10")
oct2001 <- sample_n(oct2001, 405)
oct2002 <- subset(backno1, year == "2002" & month == "10")
oct2002 <- sample_n(oct2002, 624)
oct2003 <- subset(backno1, year == "2003" & month == "10")
oct2003 <- sample_n(oct2003, 590)
oct2004 <- subset(backno1, year == "2004" & month == "10")
oct2004 <- sample_n(oct2004, 472)
oct2005 <- subset(backno1, year == "2005" & month == "10")
oct2005 <- sample_n(oct2005, 776)
oct2006 <- subset(backno1, year == "2006" & month == "10")
oct2006 <- sample_n(oct2006, 1096)
oct2007 <- subset(backno1, year == "2007" & month == "10")
oct2007 <- sample_n(oct2007, 624)
oct2008 <- subset(backno1, year == "2008" & month == "10")
oct2008 <- sample_n(oct2008, 455)
oct2009 <- subset(backno1, year == "2009" & month == "10")
oct2009 <- sample_n(oct2009, 556)
oct2010 <- subset(backno1, year == "2010" & month == "10")
oct2010 <- sample_n(oct2010, 1096)
oct2011 <- subset(backno1, year == "2011" & month == "10")
oct2011 <- sample_n(oct2011, 1062)
oct2013 <- subset(backno1, year == "2013" & month == "10")
oct2013 <- sample_n(oct2013, 119)
oct2014 <- subset(backno1, year == "2014" & month == "10")
oct2014 <- sample_n(oct2014, 51)
oct <- rbind(oct1998, oct1999, oct2000, oct2001, oct2002, oct2003, oct2004, oct2005, oct2006, oct2007, oct2008, oct2009, oct2010, oct2011, oct2013, oct2014)

Nov 1998 1123 1999 835 2000 1025 2001 495 2002 420 2003 311 2004 604 2005 858 2006 616 2007 553 2008 748 2009 915 2010 875 2011 593 2013 12 2015 17

nov1998 <- subset(backno1, year == "1998" & month == "11")
nov1998 <- sample_n(nov1998, 1123) 
nov1999 <- subset(backno1, year == "1999" & month == "11")
nov1999 <- sample_n(nov1999, 835) 
nov2000 <- subset(backno1, year == "2000" & month == "11")
nov2000 <- sample_n(nov2000, 1025) 
nov2001 <- subset(backno1, year == "2001" & month == "11")
nov2001 <- sample_n(nov2001, 495)
nov2002 <- subset(backno1, year == "2002" & month == "11")
nov2002 <- sample_n(nov2002, 420)
nov2003 <- subset(backno1, year == "2003" & month == "11")
nov2003 <- sample_n(nov2003, 311)
nov2004 <- subset(backno1, year == "2004" & month == "11")
nov2004 <- sample_n(nov2004, 604)
nov2005 <- subset(backno1, year == "2005" & month == "11")
nov2005 <- sample_n(nov2005, 858)
nov2006 <- subset(backno1, year == "2006" & month == "11")
nov2006 <- sample_n(nov2006, 616)
nov2007 <- subset(backno1, year == "2007" & month == "11")
nov2007 <- sample_n(nov2007, 553)
nov2008 <- subset(backno1, year == "2008" & month == "11")
nov2008 <- sample_n(nov2008, 748)
nov2009 <- subset(backno1, year == "2009" & month == "11")
nov2009 <- sample_n(nov2009, 915)
nov2010 <- subset(backno1, year == "2010" & month == "11")
nov2010 <- sample_n(nov2010, 875)
nov2011 <- subset(backno1, year == "2011" & month == "11")
nov2011 <- sample_n(nov2011, 593)
nov2013 <- subset(backno1, year == "2013" & month == "11")
nov2013 <- sample_n(nov2013, 12)
nov2015 <- subset(backno1, year == "2015" & month == "11")
nov2015 <- sample_n(nov2015, 17)
nov <- rbind(nov1998, nov1999, nov2000, nov2001, nov2002, nov2003, nov2004, nov2005, nov2006, nov2007, nov2008, nov2009, nov2010, nov2011, nov2013, nov2015)

Dec 1998 107 1999 551 2000 725 2001 1430 2002 966 2003 1014 2004 1101 2005 416 2006 783 2007 821 2008 589 2009 454 2010 560 2011 483

dec1998 <- subset(backno1, year == "1998" & month == "12")
dec1998 <- sample_n(dec1998, 107) 
dec1999 <- subset(backno1, year == "1999" & month == "12")
dec1999 <- sample_n(dec1999, 551) 
dec2000 <- subset(backno1, year == "2000" & month == "12")
dec2000 <- sample_n(dec2000, 725) 
dec2001 <- subset(backno1, year == "2001" & month == "12")
dec2001 <- sample_n(dec2001, 1430)
dec2002 <- subset(backno1, year == "2002" & month == "12")
dec2002 <- sample_n(dec2002, 966)
dec2003 <- subset(backno1, year == "2003" & month == "12")
dec2003 <- sample_n(dec2003, 1014)
dec2004 <- subset(backno1, year == "2004" & month == "12")
dec2004 <- sample_n(dec2004, 1101)
dec2005 <- subset(backno1, year == "2005" & month == "12")
dec2005 <- sample_n(dec2005, 416)
dec2006 <- subset(backno1, year == "2006" & month == "12")
dec2006 <- sample_n(dec2006, 783)
dec2007 <- subset(backno1, year == "2007" & month == "12")
dec2007 <- sample_n(dec2007, 821)
dec2008 <- subset(backno1, year == "2008" & month == "12")
dec2008 <- sample_n(dec2008, 589)
dec2009 <- subset(backno1, year == "2009" & month == "12")
dec2009 <- sample_n(dec2009, 454)
dec2010 <- subset(backno1, year == "2010" & month == "12")
dec2010 <- sample_n(dec2010, 560)
dec2011 <- subset(backno1, year == "2011" & month == "12")
dec2011 <- sample_n(dec2011, 483)
dec <- rbind(dec1998, dec1999, dec2000, dec2001, dec2002, dec2003, dec2004, dec2005, dec2006, dec2007, dec2008, dec2009, dec2010, dec2011)